I rotated the dichroic mirror in order not to block the green beam for next step. For easy understanding, the first picture shows how I did the rotation. Before I did the rotation, we tried for very long time to do the mode matching, but there are always two beams, and no matter how we changed the two mirrors and the position of the cavity, it did not get better, and the misalignment seems in the horizontal direction. But after I rotated the dichroic mirror for 90 degree, which I think will not effect the transmitted beam, the mode I saw now are all vertical modes. I found a mode with round shape which is likely to be the TEM00, but I am not pretty sure that is the real fundamental mode or just some coincidence that the two beams overlap.I found its peak from the spectrum, and tried to make it larger with adjustment of the two mirrors, but it did not have an obvious change. For now the misalignment is in the vertical direction and it is true that now the beam is a little bit lower than the cavity, it maybe better to higher the beam, since we have four mirrors before the cavity, it seems not very hard to higher the beam and recover it parallel to the table. Then I will check again if I got the TEM00.